Professional Documents
Culture Documents
Preface
Women walking free in the streets and children playing safely in open spaces are at the core of the
National Development Plan goal on safety and security. These basic human rights were enjoyed by
many communities in the pre-colonial South Africa and are enjoyed by some countries in the world
today. A society that is free from fear of crime is essential not only as a basic human right but also
as the foundation of economic development of a country. Evidence in this report shows that the fear
of crime limit women’s’ engagement in various daily activities.
The current report is based on the Victims of Crime Survey (VOCS) data and some administrative
data, to a lesser extent, from the World Health Organisation (WHO). The purpose of the report is to
highlight the gender impact of crime in South Africa, with particular emphasis on the impact of crime
on women. This includes the impact of crime on households headed by women. Statistical evidence
provided in this report may assist policy makers, law enforcement agencies and civil society to gauge
progress and plan interventions for protecting women as a vulnerable group.
There are two limitations that prevent the report from going deeper and provide information normally
sought by researchers, policy makers and civil society. First, the VOCS questionnaire that has been
in use for the past seven years, does not go into much detail on crimes that mostly affect women
such as sexual offence. This limitation has, to some extent, been addressed by the new survey called
Governance, Public Safety and Justice (GPSJ) which has more details on sexual offences. The first
results of the GPSJ survey will be published in August 2019. The second challenge is the effective
sample size for certain types of crime such as murder and sexual offences. While the VOCS sample
size of about 30 000 households is reasonably large, the number of cases for individuals or
households that experience these crimes is often very small – to the extent that the sampling error
for population estimates is typically very high. Although the terms murder and homicide mean
different things, we shall use the terms interchangeably in this report since we use the former during
the survey due to familiarity to the respondents, while we know that we are actually measuring the
later.
Notwithstanding the limitations mentioned in the previous paragraph, the current report still provides
useful information about crime against women. The current report is a revised version of the report
released on Tuesday June 19, 2018. The previous report was withdrawn after a stakeholder brought
to our attention the potentially misleading presentation of an estimate of female homicide rate.
………………………….
Risenga Maluleke
Statistician-General
Users may apply or process this data, provided Statistics South Africa (Stats SA) is acknowledged
as the original source of the data; that it is specified that the application and/or analysis is the result
of the user's independent processing of the data; and that neither the basic data nor any reprocessed
version or application thereof may be sold or offered for sale in any form whatsoever without prior
permission from Stats SA.
A complete set of Stats SA publications is available at Stats SA Library and the following libraries:
National Library of South Africa, Pretoria Division
National Library of South Africa, Cape Town Division
Library of Parliament, Cape Town
Bloemfontein Public Library
Natal Society Library, Pietermaritzburg
Johannesburg Public Library
Eastern Cape Library Services, King William's Town
Central Regional Library, Polokwane
Central Reference Library, Nelspruit
Central Reference Collection, Kimberley
Central Reference Library, Mmabatho
Raphael Kasonga
Email: kasonga@statssa.gov.za
www.statssa.gov.za
info@statssa.gov.za
T +27 12 336 0109
F +27 12 310 8500
CONTENTS
2. INTRODUCTION ........................................................................................................ 6
3. BACKGROUND .......................................................................................................... 7
5. CONCLUSIONS........................................................................................................ 21
7. REFERENCES ......................................................................................................... 23
9. ABBREVIATIONS ..................................................................................................... 24
List of Tables
Table 1: Quality levels based on CVs ............................................................................................ 6
Table 2: Percentage of people who think it is acceptable for a man to hit a woman, by race
and gender ....................................................................................................................... 9
Table 3: Percentage of households prevented from partaking in activities due to fear of crime . 12
Table 4: Experience of household crimes, by gender of the household head ............................. 16
Table 5: Percentage of households that experienced homicide, 2013/14-2016/17……………...17
List of Figures
Figure 1: Comparison of trends in murder rates of women, South Africa vs the world ....................7
Figure 2: Comparison of trends in men and women murder rates in South Africa, WHO 2016 ......8
Figure 3: Percentage of individuals who think it is acceptable for a husband to hit his wife, by
gender ............................................................................................................................ 10
Figure 8: Shelter and places of safety that households know about, by gender ........................... 14
Figure 10: Experience of household crimes, by gender of the houhseold head ............................. 16
1. EXECUTIVE SUMMARY
Violence against women has received increasing national and international attention accompanied
by calls for stringent measures to mitigate against this social ill. The current report provides an
overview of all types of crime, including violent crimes, as they affect women in South Africa. It seems
reasonable to conclude that, even though victimisation of women is of primary interest, the story
would not be complete without comparisons of the effect of crime on men. Hence, in almost all
instances, disaggregation according to gender was used in calculating population estimates.
Analysis of household crimes showed that male-headed households were targeted more than
female-headed households for most crimes considered in this report. In many cases a greater
percentage of male-headed households than female-headed households experienced crime. A
possible explanation that was also given in the VOCS 2016/17 release is that it may be that female-
headed households always have someone at home. The presence of people in a house may act as
a deterrent to household crime. The exception is theft of animals and possibly murder, which a
greater percentage of female-headed households experienced. Note that a female-headed
household experiencing a certain type of crime, say murder, does not mean that the victim was
female.
In the case of individual crimes, the percentage of individuals aged 16 or older who experienced any
type of crime in 2016/17 was 3,6% for women and 3,9% for men. Greater percentages of victims of
theft of personal property, sexual offence and fraud were women, while greater percentages of
victims of hijacking of motor vehicles, street robbery and assault were men. The victims of crime
survey does not produce sufficient data to provide accurate estimates of women homicide statistics.
All gender-disaggregated statistics on murder produced from VOCS have to be used with care. The
World Health Organisation (WHO) data shows that in 2000 the murder rate of women in South Africa
was more than five times the world average but it has steadily declined during the period 2000 to
2015. The data further shows that between 2000 and 2015 the murder rate for men in South Africa
has consistently been more than three times the rate for women but both rates have steadily declined
during the same period.
In terms of reporting of individual crime, greater proportions of women victims reported hijacking of
motor vehicles, sexual offence and assault. In general men were more satisfied with the police
response than women except for street robbery and assault, where greater percentages of women
were more satisfied than men.
A deeper analysis of household satisfaction with the police response to crime reporting revealed that
satisfaction is influenced more by perceptions concerning the police service than the gender of the
household head. Population group of the household head is another important factor that determines
satisfaction with police response. A white-headed household has more than four times the odds of
being satisfied with the police than a black African-headed household.
Perceptions and fear of crime are crucial impact indicators in the National Development Plan.
Evidence from the VOCS series shows that the percentage of male- and female-headed households
who thought that the level of crime increased, declined steadily between 2014/15 and 2016/17 after
an increase between 2013/14 and 2014/15. This is progress in the right direction, but the fear of
crime is still a challenge. In 2016/17, people refrained from ordinary daily activities due to fear of
crime. Women are more affected by this compared to men, as they felt they were not free to express
their sexual orientation or walk to fetch firewood or water. The pattern of feelings of safety for men is
the same as that of women, but women felt more unsafe than men. Women felt more unsafe than
men walking in their neighbourhoods alone both during the day and when it is dark.
An unexpected finding was that women had the same pattern of attitudes towards domestic violence
as men. For both men and women, the highest percentage of individuals thought it was acceptable
for a man to hit a woman if she argues with him, and the lowest percentage of individuals thought it
was acceptable for a man to hit a woman if she burns food. Black African men had the highest
percentage of individuals who thought it was acceptable for a man to hit a woman, followed by white
women.
2. INTRODUCTION
The purpose of this report is to present statistics for some types of crime perpetrated against women
in South Africa in order to provide an assessment of the trends in, and extent of victimisation of
women. Although the crime rate, in general, has decreased between 2013/14 and 2016/17[2]. violent
crimes against women, such as sexual assault, increased drastically between 2015/16 and 2016/17.
Women’s’ perception of crime is also another important area that warrants investigation. Perceptions
of whether crime is increasing or decreasing and the level of fear of crime associated with these
perceptions are important matters that need to be understood in order to inform policy. Perceptions
are so important that the first half of the NDP Vision 2030 on crime and safety is stated in terms of
perceptions and feelings of safety rather than the actual experience of crime.
“In 2030, people living in South Africa feel safe at home, at school and at work, and they enjoy a
community life free of fear. Women walk freely in the streets and children play safely outside”
Sections 4.1 – 4.2 will present results on perceptions of crime and feelings of safety based on the
Victims of Crime Survey (VOCS) data. Sections 4.3 – 4.4 deal with knowledge and accessibility of
places of safety for victims of crime, while the last four sections (4.5 – 4.8) will focus on experience
of crime, reporting of crime to the police and satisfaction with police response. Most statistics are
based on VOCS data.
Quality of estimates
In this report, every estimate will be assigned a quality level (colour) based on the coefficient of
variation of the estimate. The coefficient of variation (CV) is a measure of the relative size of error
defined as
Standard error
100
Estimate value
The survey package of the R software was used to calculate the estimates and the CVs. The package
is specifically designed for analysis of data from complex surveys. Every computation using the
survey package requires specification of three key design parameters, namely the strata, clusters
(PSUs) and final weights. The default method used for the estimation of variance is the Taylor
Linearisation technique.
3. BACKGROUND
Crime in general does not have gender preference with the exception of sexual offences which often
targets females. Gender-based violence is a topic that has received increasing attention in the
country and abroad, accompanied by calls for stringent measures to mitigate against this social ill.
Concerning the safety of women, Chapter 12 of the National Development Plan (NDP) stipulates that
“gender-based violence in South Africa is unacceptably high”. Gender-based violence is defined as
violence that targets individuals or groups of individuals on the basis of their gender. An example of
gender-based violence was the murder of Eudy Simelane in 2008 because she was a lesbian.
Femicide is a specific form of gender-based violence where women are murdered, usually by men,
just because they are female. Feminist author Diana E.H. Russell[7], an early pioneer of the term,
prefers a variant of this narrow definition of femicide namely “the killing of females by males because
they are females”. Using Russell’s narrow definition, the murder of Eudy would not count as femicide.
Later in Russell and Harmes[8] a broader definition of femicide is proposed. Briefly it is defined as
death of females resulting from any form of abuse by males. A number of other definitions of the term
have been proposed by authors, researchers and activists[9].
Despite the widespread use of the word in the media, femicide, in the narrow definition of the term,
is a rare phenomenon in South Africa, while the rate of murder of women is unacceptably high
compared to the global average.
Figure 1: Comparison of trends in murder rates of women, South Africa vs the world[1]
20
15
10
0
2000 2005 2010 2015
SA_rate Global_rate
Figure 1 shows that in 2000, the South African murder rate of women was more than five times the
global average. According to the World Health Organization (WHO) data presented in Figure 1, the
gap between South Africa and the World has consistently narrowed down between 2000 and 2015.
Africa Check[6] used the WHO female homicide data to highlight the magnitude of femicide in South
Africa compared to the world average.
The same WHO source also contained data on male homicide in South Africa. Figure 2 below
compares the trends of male and female homicide rates in South Africa between 2000 and 2015.
Figure 2: Comparison of trends in men and women murder rates in South Africa, WHO 2016[1]
60
50
Percentage
40
30
20
10
0
2000 2005 2010 2015
SA_Men 65.3 55.3 50.2 43.4
SA_Women 21.4 17.3 13.4 9.6
The homicide rate for men consistently stood at more than three times the rate for women during the
period 2000–2015. The data suggests that men are three times more likely to be victims of murder
in South Africa compared to women.
Rape, targeting women and girls, is a serious problem in South Africa. The 2016/17 Victims of Crime
statistical release[2] reported that 250 out of every 100 000 women were victims of sexual offences
compared to 120 out of every 100 000 men. Using the 2016/17 South African Police Service
statistics, in which 80% of the reported sexual offences were rape, together with Statistics South
Africa’s estimate that 68,5% of the sexual offences victims were women, we obtain a crude estimate
of the number of women raped per 100 000 as 138. This figure is among the highest in the world.
For this reason, some have labelled South Africa as the “rape capital of the world[3].
Most studies and media reports on crime against women focus on violent crimes. It is difficult to find
a study on the experiences of women on theft of personal property, street robbery, fraud and other
non-violent crimes. Although these crimes are in general not gender-based, we will present estimates
disaggregated according to gender in order to determine whether there is any gender difference in
the level of victimisation.
In this chapter we present results from an analysis of the VOCS 2016/17 data. Trends over a four-
year period (2013/14 – 2016/17) are presented only with respect to perceptions about whether crime
increased, decreased or stayed the same during the past three years. It is important to mention from
the outset, that since VOCS was not specifically designed to measure crime against women, the
scope covered by the results presented here is very limited. Many questions of interest cannot be
answered from the 2016/17 VOCS data.
The chapter begins with issues of attitudes and perceptions followed by results on fear of crime,
knowledge of shelters and places of safety for victims of crime, and accessibility of those shelters
and places of safety. The next two sections deal with the actual experience of crime by households
and individuals, respectively. The last two sections discuss reporting of crime to the police and
satisfaction with police response.
We start where it all begins, that is, in the human mind. Attitudes and perceptions play a very
important role in shaping human behaviour, including criminal activity and vulnerability to crime.
Attitudes towards women, driven mostly by cultural and religious beliefs, determine how women are
treated in society. This includes attitudes of women about themselves. A few questions were included
in the VOCS questionnaire in an attempt to capture the attitudes and perceptions of both men and
women concerning violence against women, constitutional rights and crime trends in South Africa.
The first question on attitudes asked respondents whether, in general, it is acceptable for a man to
hit a woman. It is estimated from the survey that 3,3% of men and 2,3% of women in South Africa
think it is acceptable for a man to hit a woman. It could therefore be fair to conclude that it is not
possible to eliminate violence against women when there are women who still believe that it is
acceptable to be hit by a man. Nevertheless it is encouraging to see that these percentages are small
and one would hope that they will keep on falling.
We also explored whether there was any difference in attitudes between different population groups
on the question of violence against women.
Table 2: Percentage of people who think it is acceptable for a man to hit a woman, by race and gender
The estimates where the CVs are coloured in orange should be used with caution as the error level
is too high. Black African men rank the top with the highest percentage of individuals (3,8%) who
thought that it is acceptable for a man to hit a woman. The percentage of white women (2,6%) who
think it is acceptable for a man to hit a woman is greater than that of black African women (2,5%).
This is an unexpected result because most white women live in urban areas where there is greater
exposure to human rights issues and liberal thinking.
The second question on attitudes asked respondents whether it is acceptable for a husband to hit
his wife in various situations, e.g. if she goes out without telling him, if she neglects the children, if
she argues with him, if she refuses to have sex with him, or if she burns the food or cooks the wrong
type of food. Given these possible “wrongs” committed by a wife, the percentage of those who
thought it is acceptable for a man (husband) to hit a woman (wife) increased. Figure 3 below gives a
summary of the results of this question.
Figure 3: Percentage of individuals who think it is acceptable for a husband to hit his wife, by gender
9.0
5.0
4.0
3.1
3.0 2.6
0.0
Goes out without Neglects children Argues with him Refuses to have sex Burns food
telling him
Men Women
It is interesting to note that the pattern of responses for men and women is quite similar, the only
difference being that the percentages for women are slightly lower than those for men. Both groups
think in the same way concerning the seriousness of the issues. Neglecting the children and arguing
with the husband are considered by both men and women to be the most serious issue, while burning
the food is considered to be a minor issue.
We also explored if there was any relationship between the responses to the first question and those
to the second question. For those who said, in general, it is acceptable for a man to hit a woman in
Question 1, 51,1% of men and 33,8% of women responded that it is acceptable for a husband to hit
his wife if the wife is guilty of one or more of the five “wrongs”. For those who answered ‘No’ to the
first question, the percentage of those who said it is acceptable for a husband to hit his wife if the
wife is guilty of one or more of the five “wrongs” was 10,4% for men and 9,4% for women. The
conclusion is therefore that even those who think that it is not acceptable to hit a woman, also think
that in certain circumstances a husband has the right to hit his wife.
The last question on attitudes was whether women should have the same constitutional rights as
men. It is estimated that 67% of South African men agreed with this statement, while 73% of women
agreed. These relatively low percentages raise the question as to whether respondents understood
the concept of “constitutional rights”. It is possible that some understood constitutional rights as
meaning that men and women should play the same roles. Some may not know what these
constitutional rights entail. This uncertainty may be avoided by including questions on the
respondent’s knowledge of the Constitution.
Concerning perceptions of crime, household heads were asked for their opinions on how the levels
of crime have changed during the past three years. The response options were: increased,
decreased, and stayed the same. Figure 4 below gives a summary of estimates from 2016/17 VOCS
data.
45.0
40.0
35.0
30.0
Pecentage
25.0
20.0
15.0
10.0
5.0
0.0
Increased Decreased Stayed the same
Male 37.9 27.8 33.5
Female 40.9 26.7 31.6
The patterns of opinions for male and female-headed households are similar; however, the
percentage of female-headed households who thought that crime increased was greater than that of
men. A four-year data graph on the percentage of South Africans who thought that crime had
increased was plotted to determine the trends in the perceptions of both men and women.
Figure 5: Percentage of households that think crime increased, by gender of the household head
46
44
42
Percentage
40
38
36
34
2013/14 2014/15 2015/16 2016/17
Axis Title
Male Female
Figure 5 shows that, consistently over the four-year period, the percentage of women who thought
that crime increased was greater than the corresponding percentage for men. The trends, however,
exhibit similar patterns with percentages declining between 2014/15 and 2016/17, showing that
people are becoming more positive about crime trends. Women are less positive compared to men.
The NDP asserts that the fear of crime “has consequences for women and girls and their ability to
achieve their potential in every sphere of social and productive life. Gender-based violence in all its
forms denies women and girls the opportunity to achieve equality and freedoms enshrined in the
Constitution”. In this section, we use data to explore how crime impacts the lives of women. How are
the freedoms enshrined in the Constitution compromised due to fear of crime?
Table 3: Percentage of households prevented from partaking in activities due to fear of crime
The estimates in Table 3 clearly show that in general, crime impedes the activities of women more
than it does those of men, though marginally in some cases. The difference between women and
men is more significant in the rural areas, affecting activities of keeping livestock/poultry in kraals
and walking to fetch wood or water. It is interesting to note that the activities of going to open spaces
or parks, allowing children to play freely and walking to work/town top the list for both women and
men. Almost a third of the population feel unsafe going to open spaces or parks.
Walking in your area of residence without fear of being mugged or attacked by a criminal is a basic
freedom that everyone should enjoy. In VOCS, respondents were asked how safe they would feel if
they were to walk in their neighbourhoods alone. The question was split into two questions: one dealt
with walking during the day and the other dealt with walking when it is dark. Results of these
questions are summarised in Figure 6 and Figure 7, and are disaggregated according to gender.
Women feel more unsafe than men walking alone in their neighbourhoods both during the day and
when it is dark. About 49% of women would feel very safe walking alone in their neighbourhoods
during daytime compared to 54% of men. The figures are swopped when asked about walking alone
when it is dark, where 54% of women would feel very unsafe walking alone compared to 48% of
men. The estimates here are relevant to South African individuals over the age of 16.
The next question to ask is whether, should their fears become a reality and they become victims of
crime, they know shelters and places of safety for victims of crime. Estimates from VOCS 2016/17
data show that only 9% of households in South Africa know a shelter or a place of safety for victims
of domestic violence. This is true for both male- and female-headed households. For households
that know the location of a shelter or place of safety for victims of domestic violence, Figure 8 shows
the type of institution they know.
Figure 8: Shelter and places of safety that households know about, by gender of the household head
60.0
50.0
40.0
Percentage
30.0
20.0
10.0
0.0
Religious/Traditional
State run NGO/volunteer run Other
group
Male 50.8 30.0 9.5 2.5
Female 53.2 28.2 6.7 5.4
The majority of households, over 50%, know shelters and places of safety which are state-run
institutions. The difference between male- and female-headed households in terms of the type of
institutions they know is not significant. The distribution in Figure 8 is not in line with the Department
of Justice database for shelters of abused women where, in 2012, 90% of the shelters were run by
non-profit organisations (NPOs)
Knowing where to get help and being able to get help are two very different things. How accessible
are shelters and places of safety for women who experience crime? Respondents were asked how
long on average it would take them to get to a shelter or a place of safety. Figure 9 below gives a
summary of population estimates based on the responses.
Figure 9: Time it takes to travel to a shelter or place of safety, by gender of the household head
70.0
60.0
50.0
Percentage
40.0
30.0
20.0
10.0
0.0
Less than 30
31 - 60 minutes 61 - 120 minutes More than 2 hours
minutes
Male 58.4 31.6 3.7 2.1
Female 56.9 31.6 4.8 3.3
Figure 9 shows that male- and female-headed households have equal accessibility to shelters and
places of safety for victims of domestic violence. These facilities are quite accessible to households,
as the majority, over 93%, of households would take less than one hour to get there, while almost
60% would take less than half an hour.
In this section we discuss how women experience crime at household level. It is unfortunate that
household data does not contain information about experiences of individual women and girls in the
household. The data only provides the collective experience of crime of a household. We are thus
forced to rely on the gender of the household head as a vehicle towards understanding the way
women experience crime. The types of crime covered by the household section of VOCS 2016/17
include theft of motor vehicle, housebreaking or burglary, home robbery, theft of livestock/poultry and
other animals, theft of crops planted by the household, murder, trafficking of persons (human
trafficking), theft out of motor vehicle, deliberate damaging/burning/ destruction of dwelling, motor
vehicle vandalism, and theft of bicycle.
With the exception of murder and human trafficking, these are not the type of crimes that normally
receive the attention of researchers, politicians and the media. Despite being out of the spotlight,
these crimes have a great impact on the lives and well-being of women. For example, apart from the
psychological trauma that comes with home robbery, women sometimes get raped or murdered
during a home robbery. Deliberate destruction of a dwelling may sound like a benign type of crime
but may have a great impact on women. Table 4 below provides a summary of estimates of the
percentage of households that experienced household crime during the past twelve months.
Trafficking of persons was left out because the data contained very few incidents on this crime.
The bar-chart below provides a clearer picture of the levels of household crime.
Poor estimates (marked by orange CVs in Table 4) have been left out of the discussion. Comparing
the remaining estimates, we find that female-headed households experienced a higher crime rate
compared to male-headed households only for theft of animals (livestock/poultry and other animals).
This means that male-headed households are more likely to be victims of household crimes than
female-headed households. Many possible explanations can be advanced for male dominance of
household crimes.
As remarked at the beginning of the report typically the number of cases that experience homicide
in the VOCS sample is small making it difficult to obtain accurate disaggregated estimates. Table 2
presents estimates of the proportion of households that experienced homicide disaggregated by
gender of the head of household.
Table 5: Percentage of households that experienced homicide by gender of the head, 2013/14-2016/17
According to the quality criteria used in the VOCS 2016/17 statistical release all the above estimates
would be classified as “use with care”. However under the new more strict quality criteria given in
Table 1 only half of the estimates are of acceptable quality (use with care).
0.18
0.16
0.14
0.12
0.1
0.08
0.06
0.04
2013/14 2014/15 2015/16 2016/17
Male 0.161 0.082 0.087 0.061
Female 0.126 0.171 0.107 0.157
Due to doubtful quality of the estimates in Table 5 the trends in Figure 11 should be interpreted
simply as suggestive. If the above trends reflect the reality of household experiences of homicide
then female-headed households have not shown improvement with respect to this type of crime.
In the previous section crime against women in household settings was explored. The focus is now
on crimes that affect individual women. The target population surveyed were all members of South
African households of aged 16 and above. Estimates for different types of crime disaggregated by
gender are first presented and discussed.
Estimates of the percentage of males who experienced sexual offence and percentage of females
who experienced hijacking are poor estimates and therefore, these cannot be used in the analysis.
Figure 12 depicts the prevalence of the various forms of crime experienced by individual South
Africans. The pattern of types of crime experienced by women is almost similar to that experienced
by men. Theft of personal property was the most common crime experienced by women, followed
by assault, robbery, sexual offence and fraud. However, a greater percentage of men experienced
these crimes except for possibly sexual offence and fraud.
Table 6 and Figure 12 present population estimates of proportions of individuals aged 16 and above
who experienced the various types of crimes. It would also be useful to focus on victims of crime and
estimate the proportion of victims who are women. These estimates would help give an indication of
the types of crime in which women are most vulnerable.
Assault 46,7 11
Fraud 53,8 19
Table 7 shows that women are more vulnerable to theft of personal property, sexual offence and
fraud compared to men. Sexual offence, in particular, stands out with 68,5% of the victims being
women.
Failure of victims to report crime to the police is a major challenge for crime statistics. There are
many reasons that victims do not report crime to the police. VOCS reports have consistently identified
“police would do nothing or police can do nothing” as the main reasons for not reporting crime to the
police. This section seeks to explore the crime reporting level for women for each crime experienced
by individual women.
Only four of the six pairs of estimates in Table 8 are usable statistics. The error levels for estimates
of reporting rates for sexual offence and fraud are too high. The crimes that were mostly reported by
women are hijacking of motor vehicles (over 72%) and assault (over 62%). The least reported crimes
are theft of personal property and street robbery.
Victims of crime may decide not to report to the police due to the negative perceptions they have
about the police, such as believing that the police would do nothing, or they may have had
unsatisfactory experiences with the police service previously. In this section we investigate factors
that determine satisfaction with the police, in particular comparing the extent to which women are
satisfied with the police with the extent to which men are satisfied. Are women more satisfied with
the police compared to men?
Table 9 presents comparisons for different types of crime. Unfortunately, the only meaningful
comparisons that can be made are for theft of personal property and assault because the CVs are
within acceptable limits. In the case of theft of personal property, the percentage of women who were
satisfied with the police was almost half that of men, while for assault, the percentage of women who
were satisfied with the police is almost double that of men.
A formal way to investigate factors that determine satisfaction with the police is to use a logistic
regression model. We shall consider only the example of housebreaking for the purpose of this
analysis. Therefore, the factors will be household factors such as the gender and population group
of the household head. Gender, population group and general satisfaction with the police were used
as independent variables. General satisfaction (satisfaction_1) is a variable that captures the
response to a question on satisfaction with police without reference to any crime experienced. This
means that general satisfaction is about the respondent’s perception of the police service rather than
of the experience. Fitting the model produced the following results:
The above analysis was restricted to households that reported housebreaking to the police. Table
10 shows that female-headed households were not significantly more satisfied with police response
5. CONCLUSIONS
This report provided an overview of all types of crime, including violent crimes, as they affect women
in South Africa. Although victimisation of women was of primary interest, the report would have been
incomplete if comparisons with the effect of crime on men were not made. Hence, in almost all
instances, disaggregation according to gender was used in calculating the population estimates.
Analysis of household crimes showed that for most crimes considered in this report male-headed
households were targeted more than female-headed households. In many cases a greater
percentage of male-headed households than female-headed households experienced crime. A
possible explanation that was also given in the VOCS 2016/17 release is that it may be that female-
headed households always have someone at home. The presence of people in a house may act as
a deterrent to household crime. The exception is theft of animals and possibly murder, where a
greater percentage of female-headed households experienced these crimes. Note that a female-
headed household experiencing a certain type of crime, say murder, does not mean that the victim
was female.
In the case of individual crimes, the percentage of individuals aged 16 or older who experienced any
type of crime in 2016/17 was 3,6% for women and 3,9% for men. Greater percentages of women
were victims of theft of personal property, sexual offence and fraud; greater percentages of men
were victims of hijacking of motor vehicles, street robbery and assault. In terms of reporting of
individual crime, greater proportions of women victims reported hijacking of motor vehicles, sexual
offence and assault. In general, men were more satisfied with police response than women, except
for street robbery and assault, where greater percentages of women were more satisfied than men.
A deeper analysis of household satisfaction with police response to crime reporting revealed that
satisfaction is influenced more by perceptions concerning police service than the gender of the
household head. Population group of the household head is another important factor that determines
satisfaction with police response. A white-headed household has more than four times the odds of
being satisfied with the police than a black African-headed household.
Perceptions and fear of crime are crucial impact indicators in the National Development Plan.
Evidence from the VOCS series shows that the percentage of male and female-headed households
who thought that the level of crime increased declined steadily between 2014/15 and 2016/17, after
an increase between 2013/14 and 2014/15. This is progress in the right direction, but the fear of
crime is still a challenge. In 2016/17, people refrained from ordinary daily activities due to fear of
crime. Women are more affected by this than men, as they felt not free to express their sexual
orientation or to walk to fetch firewood or water. The pattern of feelings of safety for men is the same
as that of women, but women felt more unsafe than men. Women felt more unsafe than men walking
in their neighbourhoods alone both during the day and when it is dark.
An unexpected finding was that women had the same pattern of attitudes towards domestic violence
as men. For both men and women, the highest percentage of individuals thought it was acceptable
for a man to hit a woman if she argues with him, and the lowest percentage of individuals thought it
was acceptable for a man to hit a woman if she burns the food. Black African men had the highest
percentage of individuals who thought it was acceptable for a man to hit a woman, followed by white
women.
Attitudes and beliefs are the key factors that drive crime and particularly violence against women.
Non-progressive attitudes and beliefs among the people of South Africa, including women, remain a
major challenge in fighting crime against women. Evidence provided in this report also shows that
the problem is the level of crime in the country rather than crime against women. In many crimes
(including murder) men have been more victimised than women. If crime levels decline then crime
against women will also decline. This conclusion does not suggest that there is no need for targeted
interventions against crimes that victimise women. However, femicide is a term that is often misused
by activists and the media. The killing of females simply because they are females is a rare
phenomenon in South Africa, while the incidence of female homicide resulting from domestic
violence or violence emanating from broken relationships is unacceptably high.
The report relied heavily on the VOCS data but the data does not provide detailed information on
crimes that mostly affect women. For example, the data provides information about individual
experience of sexual offence but it does not provide information about the type of sexual offence
experienced (i.e. rape, sexual assault, incest or sexual touching). This limitation has been addressed
in the new survey called Governance, Public Safety and Justice. The other challenge is the effective
sample size for certain types of crime such as murder and sexual offence. While the VOCS sample
size of about 30 000 households is reasonably large, the number of hits for individuals or households
that experienced these crimes is often very small – to the extent that the sampling error for population
estimates is typically very high. This is particularly the case when a high level of disaggregation is
desired. The small effective sample size may also be due to unwillingness on the part of respondents
to answer questions which may be viewed as very private, such as sexual offence. Increased use of
administrative data such as South African Police Service (SAPS) data to complement the survey
data is likely to be the way of addressing this challenge.
7. REFERENCES
1. Global Health Estimates 2015: Deaths by Cause, Age, Sex, by Country and by Region,
2000-2015. Geneva, World Health Organization; 2016. http://apps.who.int/violence-
info/homicide
2. Statistics South Africa, Victims of Crime Survey, Release 2016/17, September 2017.
3. Ronet Bachman, Violence against Women – A National Crime Victimization Survey Report,
1994.
4. Commission for Gender Equality, Policy Brief, Combating domestic violence against
women and children in the Western Cape by increasing access and improving the quality
of services, 2013.
5. Langa-Mlambo and Soma-Pillay, Violence against women in South Africa, 2012, Q & G
Forum, Vol. 24: 17-21.
6. Africa Check, Femicide in South Africa: 3 numbers about the murdering of women
investigated, https://africacheck.org/reports/femicide-sa-3-numbers-murdering-women-
investigated/, 2017
7. Russell, Diana E.H. “The origin and the importance of the term femicide,
http://www.dianarussell.com/origin_of_femicide.html, 2011.
8. Russell, Diana E.H. and Harmes, Roberta A., (Eds.), Femicide in global perspective New
York, Teachers College Press, 2001.
8. DEFINITIONS OF TERMS
Femicide The intentional killing of females (women or girls) because they are females
Intimate femicide The killing of women by intimate male partners
Homicide The killing of a human being by another
Murder The unlawful premeditated killing of a human being by another
9. ABBREVIATIONS
CV Coefficient of variation
NDP National Development Plan
NPO Non-profit organisation
PSU Primary sampling unit
SAPS South African Police Service
VOCS Victims of Crime Survey
WHO World Health Organization